Data Quality

Connecty AI gives you data quality at every step of the pipeline with very granular control, so AI and users always see the most reliable input and output.

Data quality rules span the full lifecycle of a query:

1. Source syncing & profiling (AI Sync Process during adding data connection)

When you connect a data source, Connecty:

• Detects available and unavailable objects (tables, views, fields)

• Automatically flags PII and sensitive columns. Read more here.

• Computes column-level statistics (nulls, distincts, ranges etc)

• Builds profiling summaries that are reused in Answers, Metric Verse, and Catalog

This early profiling becomes the foundation for downstream rules and governance.

2. Input table rules (operational checks on raw data)

In the Data Quality Rule Creator, you can attach rules to the input tables used in your SQL:

• Enforce completeness and integrity (e.g., required fields, valid ranges)

• Encode internal data governance process rules

3. Output rules per dataframe

After the query runs, Connecty lets you define quality control on the result itself:

• Output dataframe rules – attach checks to each column in the final result (e.g., thresholds, expected ranges, minimum volume, time coverage).

• Per-chart quality rules – validate the data that drives each visualization, so any chart can carry its own quality guardrails.

Connecty automatically calculates statistics per output frame to help you set sensible rule thresholds (e.g., min/max, averages, distributions). Rules can be semi-automated: you start from the suggested stats and refine the conditions that matter for your business.

4. Column level data stats

• Column-level data stats and profiling are available directly in the Column Inspector under Catalog.

• Analysts can quickly see nulls, uniqueness, ranges and distributions before trusting a column in queries or KPIs.

When to use Data Quality Rules

Use Data Quality Rules whenever:

• A query or metric becomes business-critical and should be monitored

• You want to codify tribal knowledge about data quirks or edge cases

• Governance or compliance teams need operational checks enforced consistently

• You want analysts and business users to see quality signals at the point of use, without digging into separate tools

By combining business concepts (metrics, dimensions) with operational context (quality warnings, PII flags, governance rules), Connecty AI ensures that insights are not just easy to access - but also safe, compliant, and trustworthy.